Coding With Fun
Home Docker Django Node.js Articles Python pip guide FAQ Policy

Cassandra data model

May 17, 2021 Cassandra

Table of contents

Cassandra's data model is very different from the one we typically see in RDBMS. /b10> This chapter provides an overview of how Cassandra stores data.


Cassandra databases are distributed across several machines that operate together. /b20> The outer outerst container is called a cluster. b20> For troubleshooting, each node contains a copy that, in the event of a failure, is replicated. /b20> Cassandra arranges nodes in a cluster in a circular format and assigns data to them.

Key space (Keyspace)

Keyspace is the outermost container of data in Cassandra. The basic property of a keyspace in Cassandra is -

  • Replication Factor - It is the number of computers in the cluster that will receive copies of the same data.

  • Replica placement policy - It's just a policy to put a copy on media. /b20> We have simple policies (rack-aware policies), old network topology policies (rack-aware policies), and network topology policies (data center sharing policies).

  • Column family - The keyspace is a container for a list of one or more column familyes. /b20> A column family is also a container for a collection of rows. /b21> Each row contains a sequence. /b22> Column familyes represent the structure of the data. /b23> Each key space has at least one, usually many column families.

The syntax for creating a key space is as follows -

WITH replication = {'class': 'SimpleStrategy', 'replication_factor' : 3};

The following illustration shows a diagram of the key space.

Cassandra data model

Column family

A column family is a container for an ordered collection of rows. /b20> Each row is an ordered collection of columns. /b20> The following table lists the key points for distinguishing column series from relationship database tables.

The relationship table Cassandra column family
Patterns in the relationship model are fixed. /b10> Once some columns are defined for a table, all columns must be filled with at least one empty value in each row when data is inserted.

In Cassandra, although column familyes are defined, columns are not. /b10> You can freely add any column to any column family at any time.

The relationship table defines only columns, and the user populates the table with values. In Cassandra, a table contains columns, or can be defined as a super-column family.

Cassandra column family has the following properties -

  • keys_cached - It represents the number of locations where each SSTable remains cached.

  • rows_cached - The number of rows whose entire contents will be cached in memory.

  • preload_row_cache - It specifies whether you want to prefill the row cache.

Note - Unlike the relationship table for patterns that are not fixed column familyes, Cassandra does not force a single row to own all columns.

The following illustration shows an example of a Cassandra column family.

Cassandra data model


A column is Cassandra's basic data structure and has three values, namely key or column name, value, and timestamp. /b10> The structure of the columns is given below.

Cassandra data model

Super column

The super column is a special column, so it is also a key value pair. /b10> But the super column stores the map of the child column.

Typically, column familyes are stored in a single file on disk. /b20> Therefore, in order to optimize performance, it is important to keep the columns that you might query together in the same column family, and super columns can be helpful here. H ere's the structure of the super column.

Cassandra data model

Cassandra and RDBMS data models

The following table lists the key points for distinguishing between Cassandra's data model and RDBMS's data model.

Rdbms Cassandra
RDBMS processes structured data. Cassandra processes unstructured data.
It has a fixed pattern. Cassandra has a flexible architecture.
In RDBMS, a table is an array of arrays. ROW x COLUMN In Cassandra, the table is a list of nested key value pairs. (ROW x COLUMN key x COLUMN value).
A database is the outermost container that contains data that corresponds to the application. Keyspace is the outermost container that contains the data that corresponds to the application.
The table is the entity of the database. A table or column family is an entity of key space.
Row is a single record in RDBMS. Row is a replication unit in Cassandra.
The column represents the properties of the relationship. Column is the storage unit in Cassandra.
RDBMS supports the concept of foreign keys, connections. Relationships are represented by collections.